Searching Damian Gordon. Google PageRank Damian Gordon.

75
Searching Damian Gordon

Transcript of Searching Damian Gordon. Google PageRank Damian Gordon.

SearchingDamian Gordon

Google PageRankDamian Gordon

Google Search Algorithm

• First Draft:PROGRAM GoogleCollect: NextLink <- random website; WHILE (NextLink != NULL) DO IF (No copy of this page in google collection) THEN copy this page into google collection; ENDIF; NextLink <- Next link on this page; ENDWHILE;END.

Google Search Algorithm

• First Draft:PROGRAM GoogleSearch: READ SearchString; Get First Webpage from collection; WHILE (Webpages Left to Search) DO IF (SearchString IN Current-Web-Page) THEN Put this page on the list; ENDIF; Get Next Webpage; ENDWHILE; Order the list according to PageRank;END.

Database SearchingDamian Gordon

Searching

• Oracle• DB2• MySQL• SQL Server• PostgreSQL

Searching

Array SearchingDamian Gordon

Searching

• Let’s remember our integer array from before:

Searching

44 23 42 33 16 54 34 8218 ……..… 340 1 2 3 4 5 6 7 38 39Age

Searching

• Let’s say we want to find everyone who is aged 18:

Searching: Sequential Search

Searching: Sequential Search

Searching: Sequential Search

Searching: Sequential Search

Searching: Sequential Search

Searching: Sequential Search

Searching: Sequential Search

Searching: Sequential Search

Searching: Sequential Search

Searching: Sequential Search

Searching: Sequential Search

Searching: Sequential Search

Searching: Sequential Search

Searching: Sequential Search

Searching: Sequential Search

Searching: Sequential Search

Searching: Sequential Search

Searching: Sequential Search

Searching: Sequential Search

Searching: Sequential Search

Searching: Sequential Search

Searching: Sequential Search

Searching: Sequential Search

Searching: Sequential Search

Searching: Sequential Search

Searching: Sequential Search

Searching: Sequential Search

Searching: Sequential Search

Searching: Sequential Search

Searching: Sequential Search

Searching: Sequential Search

Searching: Sequential Search

Searching: Sequential Search

Searching: Sequential Search

Searching: Sequential Search

Searching: Sequential Search

Searching: Sequential Search

Searching: Sequential Search

Searching: Sequential Search

Searching: Sequential Search

Searching: Sequential Search

Searching: Sequential Search

• This is a SEQUENTIAL SEARCH.

• If the array is 40 characters long, it will take 40 checks to complete. If the array is 1000 characters long, it will take 1000 checks to complete.

• Here’s how we could do it:

PROGRAM SequentialSearch: integer SearchValue <- 18; integer ArraySize <- 40; FOR N IN 0 TO ArraySize-1 DO IF Age[N] = SearchValue THEN PRINT “User “ N “is 18”; ENDIF; ENDFOR;END.

Searching: Sequential Search

Searching: Binary Search

• If the data is sorted, we can do a BINARY SEARCH

Searching: Binary Search

• If the data is sorted, we can do a BINARY SEARCH

16 18 23 23 33 33 34 8243 ……..… 780 1 2 3 4 5 6 7 38 39Age

Searching: Binary Search

• If the data is sorted, we can do a BINARY SEARCH

Searching: Binary Search

• If the data is sorted, we can do a BINARY SEARCH

• This means we jump to the middle of the array, if the value being searched for is less than the middle value, all we have to do is search the first half of that array.

Searching: Binary Search

• If the data is sorted, we can do a BINARY SEARCH

• This means we jump to the middle of the array, if the value being searched for is less than the middle value, all we have to do is search the first half of that array.

• We search the first half of the array in the same way, jumping to the middle of it, and repeat this.

Searching: Binary Search

Searching: Binary Search

Searching: Binary Search

Searching: Binary Search

Searching: Binary Search

Searching: Binary Search

Searching: Binary Search

Searching: Binary Search

Searching: Binary Search

Searching: Binary Search

Searching: Binary Search

• The BINARY SEARCH just takes five checks to find the right value in an array of 40 elements. For an array of 1000 elements it will take 11 checks.

• This is much faster than if we searched through all the values.

• If the data is sorted, we can do a BINARY SEARCH

PROGRAM BinarySearch: integer First <- 0; integer Last <- 40; boolean IsFound <- FALSE; WHILE First <= Last AND IsFound = FALSE DO Index = (First + Last)/2; IF Age[Index] = SearchValue THEN IsFound <- TRUE; ELSE IF Age[Index] > SearchValue THEN Last <- Index-1; ELSE First <- Index+1; ENDIF; ENDIF; ENDWHILE;END.

Searching: Binary Search

etc.